Amazing talk on a very interesting topic. I was aware of how these models evolved recently, but not of how much they gained grounds even when it comes to actual real time forecasting. This is quite something.
@@cha-vz9ls Unfortunately, I have never estimated a DSGE model. However, you don't need to know about all the intricate details of how to do it to get started. You can use Dynare with either MATLAB or Octave (like MATLAB, but open source) and look online for examples of how people do it -- hopefully, with simple DSGE models first. The rough outline is that you'll typically calibrate a few parameters, look in previous papers (for things like which priors people typically use, what kind of values they usually find, etc.), and pick some method that Dynare implements for estimation. Dynare also does maximum likelihood and GMM. I don't know if it allows for IRF matching (say, like Christiano, Eichenbaum and Evans (2005)). If you don't want to use Dynare, you'll have to get your hands really, really dirty coding very complicated stuff.
